Case studies

Different organisations. Different scales. Same approach.

From a focused pilot to an entire organisation, the methodology flexes but the principle holds: every step delivers value and unlocks what's next.

Case study

Travel & Tourism

Focused team, full stack

A travel company with disconnected tools and two distinct brands needing faster, more consistent guest service. Guest Operations was spending hours on enquiries that should take minutes.

Technology enablement
Change adoption
Started with
Workshop to surface real problems and prioritise use cases
Prompt training to build confidence from day one
Built
Custom GPTs for Guest Operations with brand-specific tone across two brands
Shared prompt libraries, refined techniques, workflow standards
They now own
Production knowledge assistant with retrieval architecture and live product data
Trained team extending capability on their own
What's next
Customer-facing AI: same knowledge, same architecture, extended to guests
Voice agent for low-touch enquiries, freeing consultants for high-value conversations

Key outcomes

  • Guest Operations handles brand-specific enquiries in minutes, not hours
  • The team owns the capability and is extending it independently
  • Architecture built to enable customer-facing AI and voice without rebuilds

Every step was designed to enable what comes after. The knowledge assistant wasn't built in isolation. The architecture, data, and team capability were all built so customer-facing AI becomes an extension, not a rebuild.

Case study

Media & Broadcasting

Marketing team, capability first

A media company whose marketing team had ChatGPT licences and were using them, each in their own way. The output was generic, off brand and written for no customer in particular. At one planning day, several teams arrived with the same idea because they had all asked the same tool the same way. Creative production was already a bottleneck, and nobody had time to spare.

Technology enablement
Change adoption
Started with
Cross-functional workshop mapping pain points across the customer lifecycle, from winning a customer to winning them back
Enthusiasts and sceptics in the same room, with a waiting list to attend
Built
A use case roadmap, every idea ranked on impact against the capability to deliver it
Three hands-on prompting sessions, from fundamentals to managing drift to building custom GPTs, with drop-in help desks between
They now own
A persona custom GPT grounded in their own customer segmentation research, and a brand guide it writes to
A repeatable method for turning source documents into GPT knowledge, so the next one is theirs to build
What's next
The roadmap's next tier: campaign strategy, AI search visibility and dynamic content
Moving the wider team from ad hoc prompting to shared, reusable tools

Key outcomes

  • Copy drafted and checked against real customer segments and the brand's tone of voice, not a generic customer
  • The team can build and maintain their own custom GPTs from their own documents
  • A prioritised roadmap that takes marketing from individual prompting to shared tools

The licences were never the constraint. Shared context was. The team were taught to build rather than handed a finished tool, and the persona GPT was the worked example: their own research and brand guide, structured as knowledge, with instructions they can change themselves. At handover they held the roadmap, the method and the first GPT built with it.

Every step delivers value. Every step unlocks what's next. The knowledge assistant wasn't the end goal. It was the foundation that made customer-facing AI and voice possible.

Ready to see what this could look like for your organisation?

Start with a conversation. No pitch decks. No pressure.

Get in touchSee how we work